HVACR distribution industry professionals — or anyone in HVACR — with questions about the refrigerant transition can Ask A2L.

Ask A2L (aska2l.com) is a website that employs artificial intelligence (AI) to answer questions about how to work with the low-GWP refrigerants, such as R-454B and R-32, that are being more widely introduced because of U.S. Environmental Protection Agency (EPA) mandates. These refrigerants, classified as A2L because they are non-toxic (“A”), mildly flammable (“2”), and have a low burning velocity (“L”), require specific protocols for their safe use, storage, and transportation.

Ask A2L was developed by Heating, Air-conditioning and Refrigeration Distributors International (HARDI) and Distro, a company that harnesses AI for use by distribution businesses.

“We wanted to get distributors the tool that they can use to empower their contractor customers to be more knowledgeable, be better prepared, answer the right questions, and make the right decisions, to provide the guidance that they need in the field,” said Allison Greene, HARDI’s marketing director, during the recent HARDI Focus conference in Dallas, Texas.

Greene and Jason Sullivan, founder and CEO of Distro, sat down on the last day of the conference to talk about AI in general and Ask A2L in particular.

Sullivan said that when he started working with HARDI, some of the more publicly familiar AI platforms were “confidently” providing bad information on refrigerants. AI needs to be properly trained, Sullivan stressed, to respond to an issue as complex as how to handle specific refrigerants.

“When you’re looking at such a nuanced, important, industry-specific issue like the A2L transition, it turns out that those kind of general tools ... don’t do a very good job, but are, in fact, kind of misleading,” Sullivan said.

That’s where HARDI’s expertise came in. The group worked with Distro to input correct information into the Ask A2L platform.

By offering Ask A2L, “we’re not replacing the intelligence that we have” at HARDI, Greene said. “Rather, we’re trying to give our people more opportunity to do the work they need to by connecting with the humans at the other end.”

“This is a tool that we’re constantly improving — working with folks here at HARDI — both from a data perspective, also from a functional perspective,” Sullivan said.

Sullivan made several other points about Ask A2L and AI, including:

• Ask A2L is available for free, but can be customized by distribution companies for a flat fee. Customization means a distributor can embed Ask A2L on its website with its own logo, add catalogs and brand-specific information, and even steer users toward particular brands. Customization also allows companies to look at chat histories and statistics about how people are using the feature.

“It can be used as a great way to better understand how to be consultative to your customers,” Sullivan said. “More generally, it can also be a great sales lead-generation tool.”

• Ask A2L can be used to generate content, whether for marketing, emails, or educational courses.

• Questions should be asked with specificity for better answers, and for the platform’s learning function, which enables it to give even more specific and accurate answers as time passes.

“Giving it more information about the specifics of where you’re located, or what kind of business you’re asking this question in reference to, can be really helpful over time,” Sullivan said. “The tool does improve, as we’ve kind of talked about, and gets more specific with its answers.”

• Companies should have clear goals for what to accomplish with AI, and keep track of key performance indicators (KPI) in order to assess progress toward those goals.

“You really have to think about things from a metrics perspective,” Sullivan said. “You should have some sort of an AI strategy.”

• Companies also need to be careful to not input sensitive information to AI tools, as that information could later be accessed by others.

“There’s kind of an extra layer of danger to you if you’re throwing your data in there, because you don’t want to be in a position where, hey, your data or your usage of the tool is being somehow used to improve the tool in a way that can be helpful to your competition,” Sullivan said, “or, you know, even worse: Just exposed directly to other folks in the market.”